Author: blass uri
Date: 01:10:32 02/29/00
Go up one level in this thread
On February 29, 2000 at 01:44:15, Ratko V Tomic wrote: >Kasparov, although the strongest human vs human player is very likely not the >strongest human vs computer player. There may eventually arise a special >category of human championships where the one with the best results against >machines is the champion. Such champion may be only a weak human GM or even >lower, but his style and a way of thinking may be just right for beating >machines. > >From personal experience, my brother who is a master scores worse than I do >against the current top micro programs, even though I was only an expert in >human play (and that was over a decade ago; I haven't played in human >competitions ever since, but only against the programs). And, of course, my >brother beats me with ease. Despite my advice to him that he shouldn't think of >a machine as a human opponent (with reasoning and common sense), and thus the >apparently "deep" positional moves don't mean the machine understands what to do >with them, once the battle is under way he still can't snap out of >antropomorphizing the program. That seems to be a common trap for players >unfamiliar with computer programming, and especially the chess programming. I >think Kasparov has the same problem, as his comments on the "quantity turning >into quality" and the "new kind of intelligence" in the article suggest. I do not think kasparov has the same problem. I remember that he won easily Deep thought when karpov won deep thought only in a rook endgames when Deep thought missed a draw. kasparov also knows to change his style against computers and he did not go for tactical complications. He is right that computes may find sometimes good positional plans by search. In order to win the computer it is productive to get positions when the plan is long enough so the computer cannot see it and I believe that kasparov tried to do it in his games but it is not so easy. I believe that kasparov is the best player also against computers. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.